Sorry for missing this one when you posted it. I believe that the Hoofnagle lab has developed clinical assays quantifying with just one transition for one peptide of a protein. To do that, however, took a lot of method development and validation.
There are a number of ways you might convince a reviewer your one peptide is representative of the protein of interest. First, I think you have indicated that the peptide is unique to the protein of interest. Next to really prove it is the peptide you think it is, you might order a stable isotope labeled reference peptide and show that it coelutes with the peak you are measuring in your sample. To do this best, you would need some amount of the type of sample you used. If the reference peptide coelutes with the peak of interest and relative ion abundance is highly similar to what you have seen (Skyline has "rdotp" for this in a sampel where both light and heavy are present and measured), then you can make a very strong argument that you were measuring that one peptide that is unique to your protein.
Beyond that, I think you would need to get into synthetic labeled proteins (e.g. 15N) and showing that spiking that in and measuring its peptides produced similar results to what you have seen, essentially that your peptide of interest produces the most intense signal. And you might again look for any endogenous signal where peptides from the heavy peptides are observed.
Unfortunately, I have recently seen a case where a group convinced themselves they were measuring a peptide with SRM, which did not stand up to close inspection with PRM. Depending on how good your evidence is with high-res PRM, in both MS1 and MS/MS that can also be pretty convincing.
The short answer to your question is that I don't think you need more than one peptide. You might even be able to convince me that you have a good biomarker for a condition when you are measuring multiple peptides and one varies but another does not, because so much can happen to a protein from splice variants to PTMs. But, you have to build up a strong, convincing case. For me, at least, you could just base that assertion on some DDA runs and broad differential statistics.
Not sure any of that is helpful, but I wanted to at least take a shot at an answer to your question. Good luck with your research. Thanks for using Skyline.